[BUG] - pixi example mouse trail on chrome (windows 8.1)

[joaomcl]

I'm using pixi v4.8.5 in my project and when I started to test it in one old notebook I noticed a bug in PIXI.mesh.Rope. Went to the official pixi examples to see how it would behave and realized the bug also happens there.

Expected Behavior

Current Behavior

Environment

• pixi.js version: 4.8.5
• Browser: CHROME 81.0.4044.92
• OS: Windows 8.1 Pro
• Graphics card: Intel HD Graphics 3000

The same thing happens in v5 with the latest pixi version (using PIXI.SimpleRope).

The bug is also present when running on Opera (Version 67.0.3575.137). It doesn't happen in Firefox (75.0).

If I use the flag forceCanves: true it works as expected, but the performance is not good. The flag legacy: true doesn't fix it.

I decided to check my chrome on webglreport.com:

Firefox receives the green OK message for webgl1 and webgl2. Opera and chrome have the same results.

System says the driver is up to date.

I also have ubuntu 16.04 running on the same PC and the bug doesn't happen there (tested latest chrome and firefox).

Is there any workaround?

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[erikdubbelboer]

I'm still encountering this issue. What fixed it for me in this demo, https://pixijs.io/examples/#/demos-advanced/mouse-trail.js was to never set the historyX and historyY to all the same value. For some reason that causes an issue with the geometry on some devices (I only have this on my android phone, other devices don't have this issue).

[brunocapdevila]

@erikdubbelboer we are having the same issue on Android v13 only. What is exactly what you mean by never setting the historyX and historyY to all the same value ?

[erikdubbelboer]

I can't reproduce it anymore, and I'm not using it anymore. But I think the solution was something like historyX.unshift(mouseposition.x + Math.random());. As long as multiple points were never exactly on top of eachother it didn't happen. Maybe it would also work to check if the value being added to historyX,historyY is the exact same as the value already before it, and if so not to add it.